Building PyTorch ignores my current Python version

[franva]

🐛 Bug

I am trying to build PyTorch on my Raspberry Pi 4(Raspbian Buster) with Python 3.6.9(virtual env), but it keeps using Python 3.7

To Reproduce

Steps to reproduce the behavior:

1. git clone https://github.com/pytorch/pytorch.git
2. git checkout v1.3.1
3. sudo python setup.py build

Expected behavior

I am expecting it copys files from python 3.6.9 virtual environment.

Environment

Collecting environment information...
PyTorch version: N/A
Is debug build: N/A
CUDA used to build PyTorch: N/A

OS: Raspbian GNU/Linux 10 (buster)
GCC version: (Raspbian 8.3.0-6+rpi1) 8.3.0
CMake version: version 3.13.4

Python version: 3.6
Is CUDA available: N/A
CUDA runtime version: Could not collect
GPU models and configuration: Could not collect
Nvidia driver version: Could not collect
cuDNN version: Could not collect

Versions of relevant libraries:
[pip] Could not collect
[conda] Could not collect

Additional context

[mrshenli]

Can you confirm that the Python version when you run python setup.py build is indeed 3.6.9?

[franva]

Can you confirm that the Python version when you run python setup.py build is indeed 3.6.9?

hi @mrshenli yes, I can confirm that.
How?

python --version

I saw 3.6.9

[mrshenli]

Marking this as "triage review" as it is not obvious to me why using PY 3.6.9 to build would load 3.7 packages.

[colesbury]

It's probably because you are running sudo.

[franva]

It's probably because you are running sudo.

hi @colesbury could you explain the reason please?

Thanks

[colesbury]

sudo can change the $PATH. Try running without sudo. If that doesn't help, check $PATH and $PYTHONPATH.